Numerical modelling of light gauge cold-formed steel compression members subjected to distortional buckling at elevated temperatures

Fire safety design of building structures has received greater attention in recent times due to continuing loss of properties and lives during fires. However, fire performance of light gauge cold-formed steel structures is not well understood despite its increased usage in buildings. Cold-formed steel compression members are susceptible to various buckling modes such as local and distortional buckling and their ultimate strength behaviour is governed by these buckling modes. Therefore a research project based on experimental and numerical studies was undertaken to investigate the distortional buckling behaviour of light gauge cold-formed steel compression members under simulated fire conditions. Lipped channel sections with and without additional lips were selected with three thicknesses of 0.6, 0.8, and 0.95 mm and both low and high strength steels (G250 and G550 steels). More than 150 compression tests were undertaken first at ambient and elevated temperatures. Finite element models of the tested compression members were then developed by including the degradation of mechanical properties with increasing temperatures. Comparison of finite element analysis and experimental results showed that the developed finite element models were capable of simulating the distortional buckling and strength behaviour at ambient and elevated temperatures up to 800 °C. The validated model was used to determine the effects of mechanical properties, geometric imperfections and residual stresses on the distortional buckling behaviour and strength of cold-formed steel columns. This paper presents the details of the numerical study and the results. It demonstrated the importance of using accurate mechanical properties at elevated temperatures in order to obtain reliable strength characteristics of cold-formed steel columns under fire conditions.

Equity in the twenty-first century

Issues of equity and inequity have always been part of employment relations and are a fundamental part of the industrial landscape. For example, in most countries in the nineteenth century and a large part of the twentieth century women and members of ethnic groups (often a minority in the workforce) were barred from certain occupations, industries or work locations, and received less pay than the dominant male ethnic group for the same work. In recent decades attention has been focused on issues of equity between groups, predominantly women and different ethnic groups in the workforce. This has been embodied in industrial legislation, for example in equal pay for women and men, and frequently in specific equity legislation. In this way a whole new area of law and associated workplace practice has developed in many countries. Historically, employment relations and industrial relations research has not examined employment issues disaggregated by gender or ethnic group. Born out of concern with conflict and regulation at the workplace, studies tended to concentrate on white, male, unionized workers in manufacturing and heavy industry (Ackers, 2002, p. 4). The influential systems model crafted by Dunlop (1958) gave rise to The discipline’s preoccupation with the ‘problem of order’ [which] ensures the invisibility of women, not only because women have generally been less successful in mobilizing around their own needs and discontents, but more profoundly because this approach identifies the employment relationship as the ultimate source of power and conflict at work (Forrest, 1993, p. 410). While ‘the system approach does not deliberately exclude gender . . . by reproducing a very narrow research approach and understanding of issues of relevance for the research, gender is in general excluded or looked on as something of peripheral interest’ (Hansen, 2002, p. 198). However, long-lived patterns of gender segregation in occupations and industries, together with discriminatory access to work and social views about women and ethnic groups in the paid workforce, mean that the employment experience of women and ethnic groups is frequently quite different to that of men in the dominant ethnic group. Since the 1980s, research into women and employment has figured in the employment relations literature, but it is often relegated to a separate category in specific articles or book chapters, with women implicitly or explicitly seen as the atypical or exceptional worker (Hansen, 2002; Wajcman, 2000). The same conclusion can be reached for other groups with different labour force patterns and employment outcomes. This chapter proposes that awareness of equity issues is central to employment relations. Like industrial relations legislation and approaches, each country will have a unique set of equity policies and legislation, reflecting their history and culture. Yet while most books on employment and industrial relations deal with issues of equity in a separate chapter (most commonly on equity for women or more recently on ‘diversity’), the reality in the workplace is that all types of legislation and policies which impact on the wages and working conditions interact, and their impact cannot be disentangled one from another. When discussing equity in workplaces in the twenty-first century we are now faced with a plethora of different terms in English. Terms used include discrimination, equity, equal opportunity, affirmative action and diversity with all its variants (workplace diversity, managing diversity, and so on). There is a lack of agreed definitions, particularly when the terms are used outside of a legislative context. This ‘shifting linguistic terrain’ (Kennedy-Dubourdieu, 2006b, p. 3) varies from country to country and changes over time even within the one country. There is frequently a division made between equity and its related concepts and the range of expressions using the term ‘diversity’ (Wilson and Iles, 1999; Thomas and Ely, 1996). These present dilemmas for practitioners and researchers due to the amount and range of ideas prevalent – and the breadth of issues that are covered when we say ‘equity and diversity in employment’. To add to these dilemmas, the literature on equity and diversity has become bifurcated: the literature on workplace diversity/management diversity appears largely in the business literature while that on equity in employment appears frequently in legal and industrial relations journals. Workplaces of the twenty-first century differ from those of the nineteenth and twentieth century not only in the way they deal with individual and group differences but also in the way they interpret what are fair and equitable outcomes for different individuals and groups. These variations are the result of a range of social conditions, legislation and workplace constraints that have influenced the development of employment equity and the management of diversity. Attempts to achieve employment equity have primarily been dealt with through legislative means, and in the last fifty years this legislation has included elements of anti-discrimination, affirmative action, and equal employment opportunity in virtually all OECD countries (Mor Barak, 2005, pp. 17–52). Established on human rights and social justice principles, this legislation is based on the premise that systemic discrimination has and/or continues to exist in the labour force and particular groups of citizens have less advantageous employment outcomes. It is based on group identity, and employment equity programmes in general apply across all workplaces and are mandatory. The more recent notions of diversity in the workplace are based on ideas coming principally from the USA in the 1980s which have spread widely in the Western world since the 1990s. Broadly speaking, diversity ideas focus on individual differences either on their own or in concert with the idea of group differences. The diversity literature is based on a business case: that is diversity is profitable in a variety of ways for business, and generally lacks a social justice or human rights justification (Burgess et al., 2009, pp. 81–2). Managing diversity is represented at the organizational level as a voluntary and local programme. This chapter discusses some major models and theories for equity and diversity. It begins by charting the history of ideas about equity in employment and then briefly discusses what is meant by equality and equity. The chapter then analyses the major debates about the ways in which equity can be achieved. The more recent ideas about diversity are then discussed, including the history of these ideas and the principles which guide this concept. The following section discusses both major frameworks of equity and diversity. The chapter then raises some ways in which insights from the equity and diversity literature can inform employment relations. Finally, the future of equity and diversity ideas is discussed.

The Aboriginal voice and the left-leaning 'Australia' (Baz Luhrmann, 2008)

Arguing that Baz Luhrmann's "Australia" (2008) is a big-budget, non-independent film espousing a left-leaning political ideology in its non-racist representations of Aborigines on film, this paper suggests the addition of a 'fourth formation' to the 1984 Moore and Muecke model is warranted. According to their theorising, racist "first formation" films promote policies of assimilation whereas "second formation" films avoid overt political statements in favour of more acceptable multicultural liberalism. Moore and Muecke's seemingly ultimate "third formation films", however, blatantly foreground the director's leftist political dogma in a necessarily low budget, independent production. "Australia", on the other hand, is an advance on the third formation because its feminised Aboriginal voice is safely backed by a colossal production budget and indicates a transformation in public perceptions of Aboriginal issues. Furthermore, this paper argues that the use of low-cost post-production techniques such as voice-over narration by racially appropriate individuals and the use of diegetic song in Australia work to ensure the positive reception of the left-leaning message regarding the Stolen Generations. With these devices Luhrmann effectively counters the claims of right-wing denialists such as Andrew Bolt and Keith Windschuttle.

Reconceiving the good life : the key to sustainable globalisation

Resource-intensive, high-carbon, Western lifestyles are frequently criticised as unsustainable and deeply unsatisfying. However, these lifestyles are still attractive to the majority of Westerners and to a high proportion of the developing world’s middle classes. This paper argues that the imminent threat of catastrophic climate change constitutes an immediate political, economic and ethical challenge for citizens of the developed world that cannot be tackled by appeals to asceticism or restraint. There can be no solution to climate change until sustainable conceptions of the good life are developed that those in the west want to live and which others might want to live. While the ultimate solution to climate change is the development of low carbon lifestyles, it is important that government initiatives, governance arrangements and economic incentives support rather than undermine that search. Like the global financial crisis, the climate change crisis also demonstrates what happens when weaknesses in national, corporate and professional governance are exacerbated by weaknesses in global governance. In tackling the latter, it is critical the mistakes now evidenced in the former are avoided – including a rethinking of carbon market and carbon tax alternatives. It is also critical that individuals must take responsibility for their actions as consumers, voters and investors.

The role of academic community in higher learning : Alternatives to a drive-thru education

The purpose of our paper is to illustrate the fundamental importance of developing academic community among first year students. We argue that a sense of academic community is of fundamental importance in combating the effects of the neo-liberal economic discourse on higher education, and that the values of higher education are incongruent with those of economic rationalism. The discursive commodification of the student, and of education itself, works against the formation of community, both within the university environment and in the wider society. We argue that, at present, the dominant discourse shaping the social practice of higher education is that of neo-liberal economics. Community values stand in opposition to the dominant discourse, and are integral to the long-term survival of a socially critical and socially responsive society. We conclude that the importance of establishing a sense of academic community during the first year of university is justified by its ultimate value to society.

Active buildings : what can we do about buildings that simply stand still?

This paper presents background of our research and result of our pilot study to find methods for convincing building users to become active building participants. We speculate this is possible by allowing and motivating users to customise and manage their own built environments. The ultimate aim of this research is to develop open, flexible and adaptive systems that bring awareness to building users to the extent they recognise spaces are for them to change rather than accept spaces are fixed and they are the ones to adapt. We argue this is possible if the architectural hardware is designed to adapt to begin with and more importantly if there are appropriate user interfaces that are designed to work with the hardware. A series of simple prototypes were made to study possibilities through making, installing and experiencing them. Ideas discussed during making and experiencing of prototypes were evaluated to generate further ideas. This method was very useful to speculate unexplored and unknown issues with respect to developing user interfaces for active buildings.

The sample complexity of pattern classification with neural networks: the size of the weights is more important than the size of the network

Sample complexity results from computational learning theory, when applied to neural network learning for pattern classification problems, suggest that for good generalization performance the number of training examples should grow at least linearly with the number of adjustable parameters in the network. Results in this paper show that if a large neural network is used for a pattern classification problem and the learning algorithm finds a network with small weights that has small squared error on the training patterns, then the generalization performance depends on the size of the weights rather than the number of weights. For example, consider a two-layer feedforward network of sigmoid units, in which the sum of the magnitudes of the weights associated with each unit is bounded by A and the input dimension is n. We show that the misclassification probability is no more than a certain error estimate (that is related to squared error on the training set) plus A3 √((log n)/m) (ignoring log A and log m factors), where m is the number of training patterns. This may explain the generalization performance of neural networks, particularly when the number of training examples is considerably smaller than the number of weights. It also supports heuristics (such as weight decay and early stopping) that attempt to keep the weights small during training. The proof techniques appear to be useful for the analysis of other pattern classifiers: when the input domain is a totally bounded metric space, we use the same approach to give upper bounds on misclassification probability for classifiers with decision boundaries that are far from the training examples.

Empirical minimization

We investigate the behavior of the empirical minimization algorithm using various methods. We first analyze it by comparing the empirical, random, structure and the original one on the class, either in an additive sense, via the uniform law of large numbers, or in a multiplicative sense, using isomorphic coordinate projections. We then show that a direct analysis of the empirical minimization algorithm yields a significantly better bound, and that the estimates we obtain are essentially sharp. The method of proof we use is based on Talagrand’s concentration inequality for empirical processes.

Fast rates for estimation error and oracle inequalities for model section

We consider complexity penalization methods for model selection. These methods aim to choose a model to optimally trade off estimation and approximation errors by minimizing the sum of an empirical risk term and a complexity penalty. It is well known that if we use a bound on the maximal deviation between empirical and true risks as a complexity penalty, then the risk of our choice is no more than the approximation error plus twice the complexity penalty. There are many cases, however, where complexity penalties like this give loose upper bounds on the estimation error. In particular, if we choose a function from a suitably simple convex function class with a strictly convex loss function, then the estimation error (the difference between the risk of the empirical risk minimizer and the minimal risk in the class) approaches zero at a faster rate than the maximal deviation between empirical and true risks. In this paper, we address the question of whether it is possible to design a complexity penalized model selection method for these situations. We show that, provided the sequence of models is ordered by inclusion, in these cases we can use tight upper bounds on estimation error as a complexity penalty. Surprisingly, this is the case even in situations when the difference between the empirical risk and true risk (and indeed the error of any estimate of the approximation error) decreases much more slowly than the complexity penalty. We give an oracle inequality showing that the resulting model selection method chooses a function with risk no more than the approximation error plus a constant times the complexity penalty.

On the optimality of sample-based estimates of the expectation of the empirical minimizer

We study sample-based estimates of the expectation of the function produced by the empirical minimization algorithm. We investigate the extent to which one can estimate the rate of convergence of the empirical minimizer in a data dependent manner. We establish three main results. First, we provide an algorithm that upper bounds the expectation of the empirical minimizer in a completely data-dependent manner. This bound is based on a structural result due to Bartlett and Mendelson, which relates expectations to sample averages. Second, we show that these structural upper bounds can be loose, compared to previous bounds. In particular, we demonstrate a class for which the expectation of the empirical minimizer decreases as O(1/n) for sample size n, although the upper bound based on structural properties is Ω(1). Third, we show that this looseness of the bound is inevitable: we present an example that shows that a sharp bound cannot be universally recovered from empirical data.

Open problems in the security of learning

Machine learning has become a valuable tool for detecting and preventing malicious activity. However, as more applications employ machine learning techniques in adversarial decision-making situations, increasingly powerful attacks become possible against machine learning systems. In this paper, we present three broad research directions towards the end of developing truly secure learning. First, we suggest that finding bounds on adversarial influence is important to understand the limits of what an attacker can and cannot do to a learning system. Second, we investigate the value of adversarial capabilities-the success of an attack depends largely on what types of information and influence the attacker has. Finally, we propose directions in technologies for secure learning and suggest lines of investigation into secure techniques for learning in adversarial environments. We intend this paper to foster discussion about the security of machine learning, and we believe that the research directions we propose represent the most important directions to pursue in the quest for secure learning.

Local Rademacher complexities

We propose new bounds on the error of learning algorithms in terms of a data-dependent notion of complexity. The estimates we establish give optimal rates and are based on a local and empirical version of Rademacher averages, in the sense that the Rademacher averages are computed from the data, on a subset of functions with small empirical error. We present some applications to classification and prediction with convex function classes, and with kernel classes in particular.

Exponentiated gradient algorithms for conditional random fields and max-margin Markov Networks

Log-linear and maximum-margin models are two commonly-used methods in supervised machine learning, and are frequently used in structured prediction problems. Efficient learning of parameters in these models is therefore an important problem, and becomes a key factor when learning from very large data sets. This paper describes exponentiated gradient (EG) algorithms for training such models, where EG updates are applied to the convex dual of either the log-linear or max-margin objective function; the dual in both the log-linear and max-margin cases corresponds to minimizing a convex function with simplex constraints. We study both batch and online variants of the algorithm, and provide rates of convergence for both cases. In the max-margin case, O(1/ε) EG updates are required to reach a given accuracy ε in the dual; in contrast, for log-linear models only O(log(1/ε)) updates are required. For both the max-margin and log-linear cases, our bounds suggest that the online EG algorithm requires a factor of n less computation to reach a desired accuracy than the batch EG algorithm, where n is the number of training examples. Our experiments confirm that the online algorithms are much faster than the batch algorithms in practice. We describe how the EG updates factor in a convenient way for structured prediction problems, allowing the algorithms to be efficiently applied to problems such as sequence learning or natural language parsing. We perform extensive evaluation of the algorithms, comparing them to L-BFGS and stochastic gradient descent for log-linear models, and to SVM-Struct for max-margin models. The algorithms are applied to a multi-class problem as well as to a more complex large-scale parsing task. In all these settings, the EG algorithms presented here outperform the other methods.

Implicit online learning

Online learning algorithms have recently risen to prominence due to their strong theoretical guarantees and an increasing number of practical applications for large-scale data analysis problems. In this paper, we analyze a class of online learning algorithms based on fixed potentials and nonlinearized losses, which yields algorithms with implicit update rules. We show how to efficiently compute these updates, and we prove regret bounds for the algorithms. We apply our formulation to several special cases where our approach has benefits over existing online learning methods. In particular, we provide improved algorithms and bounds for the online metric learning problem, and show improved robustness for online linear prediction problems. Results over a variety of data sets demonstrate the advantages of our framework.

Closing the gap between bandit and full-information online optimization : high-probability regret bound

We demonstrate a modification of the algorithm of Dani et al for the online linear optimization problem in the bandit setting, which allows us to achieve an O( \sqrt{T ln T} ) regret bound in high probability against an adaptive adversary, as opposed to the in expectation result against an oblivious adversary of Dani et al. We obtain the same dependence on the dimension as that exhibited by Dani et al. The results of this paper rest firmly on those of Dani et al and the remarkable technique of Auer et al for obtaining high-probability bounds via optimistic estimates. This paper answers an open question: it eliminates the gap between the high-probability bounds obtained in the full-information vs bandit settings.